用复合函数回归式估测延边黄牛幼牛正常生长发育曲线研究

本研究是应用复合函数回归式y^=a b(x-x0)/[k kb(x-x0)]对延边黄牛24月龄体重、体尺的平均值和上下限共66项实测资料计算回归式和各月龄估测值,结果公、母牛体重实侧值与估测值之间的Σ(y-y^)2分别为286.56、176.24,拐点偏差为0.7、0.3个月龄,比应用y^=k/(1 aebx)计算的Σ(y-y^)2低5214.23、4007.05;计算公、母牛体尺资料平均值的估测值与实测值的Σ(y-y^)2,除母牛胸宽平均值外都比应用=a bx cxk计算结果低,与实测值的吻合度较高并能克服二次回归式和布劳迪回归式所存在的缺点.

Predicting the normal growth curve of calves of Yanbian yellow cattle using compound regression function

the compound regression equation (y=a+b(x-x0)....) was fitted to estimate the body sizes using 66 actual measurements of average body weight and sizes of Yanbian yellow cattle at the age of 24 mo. The results showed that the square of differences between measurements and estimation and the deviation of the point of inflection for male and female calves were 286.56, 0.7 mo and 176.24, 0.3 mo, respectively, the square of differences between measurements and estimation using compound regression equation for male and female was 5214.23, 4007.05 less than that of (y=k/(1+ae.....); the standard error of compound regression equation for the prediction of body measurements was lower than that of (y=a+bx+cx..) except chest width of cows, which fitted well to the actual measurements and eliminate the biases of quadratic polynomial and Blaudi regression equation.